American Journal of Emergency Medicine (2013) 31, 282-290

Original Contribution

Different impact of the appropriateness of Empirical antibiotics for bacteremia among younger adults and the elderly in the ED

Ching-Chi Lee MD ^a, Chia-Ming Chang MD ^a, Ming-Yuan Hong MD ^b,

Hsiang-Chin Hsu MD ^b, Wen-Chien Ko MD ^a,c,?

^aDepartment of Internal Medicine, National Cheng Kung University Hospital, Tainan 704, Taiwan ^bDepartment of Emergency Medicine, National Cheng Kung University Hospital, Tainan 704, Taiwan ^cDepartment of Medicine, Medical College, National Cheng Kung University, Tainan 704, Taiwan

Received 4 May 2012; revised 20 June 2012; accepted 22 July 2012

Abstract

Objectives: To investigate the clinical impact of age on bacteremia among adults visiting the emergency department (ED).

Methods: bacteremic adults visiting the ED from January 2008 to December 2008 were identified retrospectively. Demographic characteristics, severity, bacteremic pathogens with in vitro suscept- ibility, antimicrobial agents, and outcomes determined from chart records were analyzed as a case- control study.

Results: Of 518 eligible bacteremic adults, 288 (55.6%) elderly patients (>= 65 years old) were case patients and 230 younger patients (b 65 years) were regarded as control patients. The 28-day mortality rate was higher in the case patients than that in the control patients (11.8% vs 6.1%, P = .02). The proportion of inappropriate empirical antibiotic therapy between the survivors and nonsurvivors was similar in control patients (69.4% vs 64.3%, P = .77); but for the case patients, the proportion of inappropriate empirical antibiotic therapy in the survivors was lower than that in the non-survivors (27.6% vs 44.1%, P = .04). Of note, inappropriate empirical antibiotic therapy was also one of independent risk factors of 28-day mortality by the multivariate analyses in the case patients (odds ratio [OR] 3.65; P = .049). Other independent predictors of 28-day mortality in case patients included a high Pittsburgh bacteremia score (>= 4 points; OR 22.16; P b .001), bacteremia due to foci other than urinary tract infection (OR 9.07; P = .002), malignancy (OR 10.87; P b .001), coronary artery disease (OR 5.68; P = .01), and high serum creatinine (N 1.5 mg/dL; OR 3.44; P = .04).

Conclusions: For bacteremic adults, this study demonstrated the impact of inappropriate empirical antibiotic therapy on patients’ outcome in the elderly was greater than that in the younger adults.

(C) 2013

* Corresponding author. Department of Internal Medicine, National Cheng Kung University Hospital, 70403, Tainan, Taiwan. Tel.: +886 6 2353535×3596; fax: +886 6 2752038.

E-mail address: [email protected] (W.-C. Ko).

Introduction

Bacteremia is a serious, life-threatening condition that is associated with high morbidity and mortality, with Inhospital mortality rates of up to 30% [1,2]. Early administration of

0735-6757/$ - see front matter (C) 2013 http://dx.doi.org/10.1016/j.ajem.2012.07.024

appropriate empiric antibiotic therapy has repeatedly been shown to decrease mortality in patients with community- acquired and nosocomial Bloodstream infections [3-5]. Aging is associated with significant changes in the adaptive humoral and cell-mediated immunity [6,7], and poor cytokine re- sponses to sepsis have been discovered in the elderly [8]. Therefore, a well-recognized increase in mortality among the elderly with bacteremia has been demonstrated in previous investigations [9,10].

Apart from the ageing of the population, the increased prevalence of chronic-degenerative diseases, with suscepti- bility to frequent exacerbations, make the elderly frequent visitors of emergency department (ED) [11]. Focusing on the elderly with bacteremia, appropriate antibiotic therapy may reduce the mortality [12]. Therefore, for ED clinicians, empirical antimicrobial therapy has suggested to be initiated very often and soon among the elderly with suspected infections [7], either because of delayed diagnosis due to atypically clinical presentation [7,13] or impossibility of Diagnostic procedure [14]. However, a study emphasized the importance of appropriate empiric antibiotic therapy in elderly patients with bacteremia in the ED has not been reported in the English literature. Thus, we conduct this study to demonstrate the importance of appropriate Empiric therapy in the elderly with bacteremia by comparing the impact of inappropriate empiric therapy on the clinical outcomes of elderly and young adult patients with bacteremia in the ED.

Materials and methods

Study design and population

A retrospective case-control study was conducted between January 2008 and December 2008 at a medical center of approximately 1200 beds in southern Taiwan, and 67,585 annual visits to the emergency department (ED) were recorded during the study period, whereas the degree of ED over- crowding was calculated routinely every eight hours using the National Emergency Department Overcrowding Study (NEDOCS) score [15]. The local institutional review board has approved this study. The retrospective analysis included the records of adults who visited the ED during the study period and had significant blood cultures detected at the study ED. However, the study excluded the patients with contam- inated blood cultures or bacteremia diagnosed prior to visiting the ED. For the eligible bacteremic adults visiting the ED, elderly patients (aged >= 65) were regarded as the case group and younger adults (b 65 years old) as control patients. The medical records were reviewed for demographic information, clinical characteristics, Pittsburgh bacteremia severity scores, comorbidities, and laboratory data during the ED stay. The duration and type of antimicrobial agents, microbiological results, In vitro susceptibility, source of bacteremia, and further hospitalization were also derived from the chart records. As in previous descriptions [16], multiple bacteremic episodes in a

single patient were considered distinct events, if separated by at least seven days or different causes were listed for the respective ED visits.

The primary outcome was the overall mortality in 28 days after visiting the ED. If the patients discharged within 28 days of the ED arrival and were not followed-up at our hospital, the outcome information was retrieved by the telephone. The patients who could not be reached by the telephone were excluded.

Microbiological studies

Nurses performed the blood sampling for at two sets of blood cultures. Each set of blood samples routinely consisted of one bottle for aerobic culture and another for anaerobic culture, with approximately 5 to 8 mL of blood for each bottle. After the collection, the culture bottles were transported immediately to the clinical laboratory department, loaded into the BACTEC 9240 system (Becton Dickinson and Company, Franklin Lakes, NJ), and incubated for 5 days or until the instrument detected bacterial growth. The culture bottles that exhibited bacterial growth were Gram-stained, and the contents of the bottles were sub-cultured onto the plates with blood agar (BD; trypticase soy agar II 5% sheep blood), levine eosin-methylene BD, chocolate agar, and Centers for Disease Control and Prevention anaerobic BD for further identification. Biochemical tests and automatic identification systems were used for the final identification. Isolates were tested for in vitro susceptibility using the Kirby-Bauer method on Mueller-Hinton agar. The susceptibility results were interpreted according to the Clinical Laboratory Standard Institute guidelines [17].

Definitions

Elderly patients (case patients) were defined as the patients with 65 years or older, whereas adults younger than 65 years were referred to as younger adults (control patients). Bac- teremia was defined as 2 separate blood cultures growing the same microorganism or as a single blood culture associated either with a clinically identified source growing the same microorganism(s). Blood culture samples with potential contaminating pathogens (such as coagulase-negative Sta- phylococcus, Micrococcus, Propionebacterium acnes, Pep- tostreptococcus, or Bacillus species) were considered to be contaminated in accordance with the previously described criteria [18]. Polymicrobial bacteremia was defined as the isolation of more than 1 microbial species from a single bacteremic episode. Appropriate empiric antibiotic therapy was defined as the patients during the ED stay treated by at least a drug to which the isolated pathogen showed in vitro susceptibility. Otherwise was referred to as inappropriate empiric antibiotic therapy.

As previously described [19], community-acquired bac- teremia was found in patients visiting the ED from home

without recent hospitalization within the past 30 days, an invasive procedure either immediately before or at the time of admission, receiving long-term dialysis, and/or having intravascular devices at the time of ED visits. The severity of the bloodstream infection at the time of blood sampling was assessed using the Pittsburgh bacteremia score, a validated scoring system based on vital signs, mental status, mechanical ventilation, and the presence of cardiac arrest [20]. Underlying malignancy includes both Hematological malignancies and solid tumors, whereas the definitions of comorbidities were as previously described [21]. The sources of bacteremia were classified as lower respiratory tract infections, urinary tract infections, skin and soft-tissue infections, intra-Abdominal infections, or primary bacteremia according to the definitions from the Centers for Disease Control and Prevention [22].

Data analysis

Statistical analyses were performed using SPSS for Windows (SPSS, Chicago, IL), version 15.0. Continuous variables were expressed as the means +- SDs and compared using Student t test. Categorical variables, expressed as numbers and percentages, were compared using a ?2 test or Fisher exact test. Kaplan-Meier curves were used for the 28-day survival analysis. All significant variables (P b .05) in the univariate analysis were incorporated into a hierarchical logistic regression model. P b .05 was considered to be statistically significant.

Results

Demographics and clinical characteristics of all bacteremic patients

During the 1-year period, blood culture samples have been collected from 14,482 patients, and 813 (5.6%) patients had bacterial growth in their respective blood cultures, whereas the mean (+- SD) of the hospital-based NEDOS score was

142.8 (+- 26.2), reaching the level of “severely overcrowded.” After excluding 273 patients with contaminated blood culture samples, 7 pediatric visitors, 13 patients with bacteremia diagnosed prior to visiting the ED, and 2 patients who were lost to follow-up within 28 days of the ED arrival, 518 eligible bacteremic adults were included. Elder patients (case patients) accounted for 55.6% (288) of all study patients, and 230 adults younger than 65 years were regarded as biological data in the case an”>control patients. Of the total 518 bacteremic adults, most (456, 88.0%) patients visited the ED from the community, but 25 (4.8%) patients were transferred from long-term health- care facilities, and 37 (7.1%) were transferred from other hospitals. Comorbidities included hypertension (239 patients, 46.1%), diabetes mellitus (190, 36.7%), malignancy (112,

21.6%), Chronic renal insufficiency (85, 16.4%), liver

cirrhosis (69, 13.3%), coronary artery disease (43, 8.3%), congestive heart failure (36, 6.9%), history of stroke (34, 6.6%), and chronic obstructive pulmonary disease (24, 4.6%).

Most patients (433, 83.6%) were admitted to general wards and 68 (13.1%) to Intensive care units . There were 17 (3.3%) patients discharged from the ED. Fourteen patients died during their ED stay; 34 hospitalized patients died, and no mortality after ED discharge was observed within 28 days of their ED arrival. Therefore, the 28-day mortality was 9.3% (48 patients). Major sources of bacteremia included urinary tract infections (182 patients, 35.1%), primary bacteremia (85, 16.4%), pneumonia (60,

11.6%), biliary tract infections (45, 8.7%), skin and soft-

tissue infections (44, 8.5%), Liver abscess (27, 5.2%), intra-

abdominal infections (22, 4.2%), Infective endocarditis (22, 4.2%), vascular catheter infections (18, 3.5%), bone and joint infections (7, 1.4%), nose and throat infections (4, 0.8%), and central nerve system infections (2, 0.4%).

Clinical characteristics, empirical antibiotics, and outcomes in the case and control patients

Clinical variables including demographics, sources of bacteremia, severity, and comorbidities were compared between the case and control patients, as shown in Table 1. Being a resident in long-term healthcare facilities, severe sepsis (a high Pittsburgh bacteremia score >= 4 points), bacteremia due to pneumonia, having at least one comorbid- ities, hypertension, coronary artery diseases, congestive heart failure, old stroke, and chronic obstructive pulmonary disease were more often present in case patients, whereas infective endocarditis was more often noted in control patients.

Clinical outcome, appropriateness, and types of empiric antimicrobial therapy were listed in Table 2. About 70% of the patients with bloodstream infections received appropriate empirical antimicrobial agents for their bacteremic patho- gens, irrespective of case patients or control patients. Of note, the interval between the arrival of ED and the administration of antibiotics was short, in average less than

1.5 hours. More aminopenicillin/?-lactamase inhibitors or second generation cephalosporins were prescribed for case patients and more first generation cephalosporins for control patients (Table 2). The combinations of an aminoglycoside and a ?-lactam agent were uncommonly given as empirical therapy in the ED. Not surprisingly, the 28-day mortality rate in control patients (14/230, 6.1%) was lower than that in case patients (34/288, 11.8%; P = .02).

microbiological data in the case and control patients

Since there were 51 episodes of polymicrobial bacter- emia, a total of 574 isolates were obtained from the 518 eligible patients (Table 3). These pathogens were mainly

Gender, male	147 (51.0)	109 (47.4)	.40
Residents in long-term	22 (7.6)	3 (1.3)	.001
healthcare facilities Bacteremic features
Polymicrobial 28 (9.7)		23 (10.0)	.91
Community-acquired 215 (74.7)		177 (77.0)	.54
infections Severity of illness in the ED Pittsburgh bacteremia 65 (22.6)		31 (13.5)	.008
Admitted to ICUs 45 (15.6) through the ED Source of bacteremia		23 (10.0)	.06
Urinary tract infections 106 (36.8)		76 (33.0)	.37
Primary bacteremia	44 (15.3)	41 (17.8)	.43
Pneumonia	42 (14.6)	18 (7.8)	.02
Biliary tract infections	31 (10.8)	14 (6.1)	.06
Skin and soft-tissue	22 (7.6)	22 (9.6)	.43
infections Liver abscess	15 (5.2)	12 (5.2)	.99
Intra-abdomen	10 (3.5)	12 (5.2)	.32
infections Vascular catheter-	7 (2.4)	11 (4.8)	.14
related infections
Infective endocarditis	5 (1.7)	17 (7.4)	.002
Bone and joint	4 (1.4)	3 (1.3)	.93

Gram-negative aerobes (432 isolates, 75.3%) and Gram- positive aerobes (155, 27.0%). Escherichia coli (226,

Table 1 Demographic data and clinical features of bacteremia in the elderly (case patients) and the younger adults (control patients) in an emergency department (ED)

Clinical characters Patient number (%) P

Case patients, Control patients, values n = 288 n = 230

infections

score >= 4 points

pulmonary disease

ICUs = intensive care units.

infections Nose and throat infections Central nervous	1 (0.3) 1 (0.3)	3 (1.3) 1 (0.4)	.21 1.00	Inappropriate therapy Type of antibiotics First generation cephalosporin	85 (29.5) 85 (29.5)	71 (30.9) 99 (43.0)	.73 .001
system infections Third-generation 59 (20.5)						48 (20.9)	.91
Comorbidities N 1 comorbidities	185 (64.2)	107 (46.5)	b .001	cephalosporins Aminopenicillin/BLI	59 (20.5)	25 (10.9)	.003
Hypertension	164 (56.9)	75 (32.6)	b .001	Second generation	44 (15.3)	17 (7.4)	.006
Diabetes mellitus	107 (37.2)	83 (36.1)	.80	cephalosporin
Malignancy	61 (21.2)	51 (22.2)	.78	Fourth-generation	12 (4.2)	6 (2.6)	.34
Chronic renal insufficiency	47 (16.3)	38 (16.5)	.95	cephalosporins Ureidopenicillin/BLI	6 (2.1)	4 (1.7)	1.00
Coronary artery diseases	36 (12.5)	7 (3.0)	b .001	Fluoroquinolones Glycopeptides	5 (1.7) 3 (1.0)	8 (3.5) 2 (0.9)	.21 1.00
Liver cirrhosis	32 (11.1)	37 (16.1)	.09	Ureidopenicillins	3 (1.0)	5 (2.2)	.48
Congestive heart	26 (9.0)	10 (4.3)	.03	Penicillin G	0 (0)	2 (0.9)	.20
failure Old stroke	26 (9.0)	8 (3.5)	.01	Combination with an aminoglycoside	9 (3.1)	9 (3.9)	.62
Chronic obstructive	22 (7.6)	1 (0.4)	b .001 None 12 (4.2) 14 (6.1) .32

39.4%) and Klebsiella species (98, 17.1%) were two major

Gram-negative aerobes, and Staphylococcus aureus (58, 10.1%) was the major Gram-positive aerobes. Only 7 anaerobes (Clostridium species, 3 isolates; Bacteroides species, 4) were identified. The distribution of major bacterial species was similar between the case and control patients, except that S aureus was more often isolated from control patients (Table 3).

Comparisons of in vitro susceptibility between the case and control patients

In vitro susceptibilities of commonly used antibiotics for major pathogens were shown in Table 4: cefazolin, ceftriaxone, or cefpirome for Enterobacteriaceae; piper- acillin, ceftazidime, or cefepime for Pseudomonas aerugi- nosa; oxacillin for S aureus; and ampicillin for enterococci. The susceptibility rate among the above pathogens was similar, except that more methicillin-resistant S aureus was noted in case patients than control patients (15/24, 62.5% vs 10/34, 29.4%; P = .01). The phenotype of extended- spectrum ?-lactamase production was present in 4.4% to

Table 2 Empirical antimicrobial therapy and clinical outcome for bacteremia in the elderly (case patients) and the younger adults (control patients) in an emergency department (ED)

Characters Patient number (%) P Case patients, Control patients, values n = 288 n = 230

SD = standard deviation; BLI = ?-lactamase inhibitor.

.26

.02

29.8 +- 10.1

14 (6.1)

30.4 +- 11.8

34 (11.8)

.24

1.3 +- 6.2

1.0 +- 4.1

time to antibiotic administration Length of stay Crude mortality

at 28 days

Duration in the ED (hours), mean +- SD

Empiric antibiotic therapy in the ED

Gram-negative	235 (75.8)	177 (67.0)	432 (75.3)
aerobes Escherichia coli	128 (41.3)	98 (37.1)	226 (39.4)
Klebsiella species	53 (17.1)	45 (17.0)	98 (17.1)
Pseudomonas	11 (3.5)	5 (1.9)	16 (2.8)
aeruginosa Proteus species	9 (2.9)	3 (1.1)	12 (2.1)
Morganella	7 (2.3)	3 (1.1)	10 (1.7)
morganii Enterobacter	5 (1.6)	7 (2.7)	12 (2.1)
cloacae
Others	22 (7.1)	15 (5.7)	37 (6.4)
Gram-positive aerobes	73 (23.5)	82 (31.1)	155 (27.0)
Staphylococci ?	27 (8.7)	35 (13.3)	62 (10.8)
Staphylococcus	24 (7.7)	34 (12.9)	58 (10.1)
aureus ? Streptococci	30 (9.7)	41 (15.5)	71 (12.4)
Streptococcus viridans	9 (2.9)	7 (2.7)	16 (2.8)
Streptococcus	5 (1.6)	6 (2.3)	11 (1.9)
pneumoniae Streptococcus	4 (1.3)	3 (1.1)	7 (1.2)
bovis Streptococcus	4 (1.3)	2 (0.8)	6 (1.0)
pyogenes Streptococcus	4 (1.3)	2 (0.8)	6 (1.0)
agalactiae Streptococcus	2 (0.6)	4 (1.5)	6 (1.0)
intermedius Enterococci	15 (4.8)	6 (2.3)	21 (3.7)
Anaerobes	2 (0.6)	5 (1.9)	7 (1.2)
* The differences in the proportions of bacteremic isolates between the elderly and young adults were significant (P b .05).

5.7% of E coli or K pneumoniae bacteremic isolates from the case or control patients.

Table 3 Microbiologic results of 574 pathogens in 518 bacteremic adults

Bacteremic isolates

Isolate number (%)

Case patients, Control patients, Total No.

n = 310 n = 264 (%)

Risk factors of 28-day mortality among control patients

The association of several clinical variables and 28-day mortality was examined in Table 5. Male gender, underlying malignancy, high Pittsburgh bacteremia scores (>= 4 points) and a high BUN (N 20 mg/dL) in the ED were significantly associated with the 28-day mortality. In contrast, bacteremia due to a urinary tract infection heralded a better outcome. Of note, the appropriateness of empiric antibiotic therapy administrated in the ED was not significantly related to the 28-day mortality (inappropriate: 9/159, 5.7% vs appropriate:

5/72, 6.9%; P = .77). By the multivariate analysis, there was only one significant variable, a Pittsburgh bacteremia score

>= 4 points (odds ratio [OR] 13.7, 95% confidence interval [CI] 3.3-57.1; P b .001).

Risk factors of 28-day mortality among case patients

The variables significantly associated with the 28-day mortality included male gender, admission to intensive care units, underlying malignancy or coronary artery disease, bacteremic pneumonia, a Pittsburgh bacteremia score >= 4 points, serum creatinine N 1.5 mg/dL, BUN N 20 mg/dL, and C-reactive protein N 100 mg/L in the ED (Table 6). Moreover, community-acquired bacteremia and bacteremia due to a urinary tract infection would result in a favorable outcome at 28 days. Of note, a higher 28-day mortality rate was noted in case patients treated with inappropriate Empiric antibiotics than those treated with appropriate empiric antibiotics (15/85, 17.6%, vs. 19/203, 9.4%; P = .047). By the multivariate analysis, 6 variables including a Pittsburgh bacteremia score >= 4 points (OR 22.2, 95% CI 6.7-73.3; P b

.001), malignancy (OR 10.9, 95% CI 3.1-38.6; P b .001),

bacteremia due to foci other than urinary tract infection (OR 9.1, 95% CI 2.1-38.7; P = .002), coronary artery disease (OR 5.7, 95% CI 1.5-21.5; P = .01), inappropriate empiric

R = resistant; ESBL = extended-spectrum ?-lactamase.

2/6 (33.3) 1.00

4/15 (26.7)

0/5 (0)

0/5 (0)

0/5 (0)

10/34 (29.4) .01

1/11 (9.1)

1/11 (9.1)

1/11 (9.1)

15/24 (62.5)

Piperacillin-R Ceftazidime-R Cefepime-R Methicillin-R,

Staphylococcus aureus, n = 58

Ampicillin-R, enterococci, n = 21

Pseudomonas aeruginosa, n = 16 1.00

4/45 (8.9)

3/45 (6.7)

2/45 (4.4)

5/53 (9.4)

3/53 (5.7)

3/53 (5.7)

.80

20/98 (20.4)

8/98 (8.1)

5/98 (5.1)

25/128 (19.5)

10/128 (7.8)

6/128 (4.7)

1.00

Escherichia coli, n = 226 Cefazolin-R Ceftriaxone-R

ESBL-producers Klebsiella species, n = 98 Cefazolin-R

Ceftriaxone-R ESBL-producers

Table 4 A comparison of in vitro resistance (to commonly prescribed antibiotics in the ED) of the common bacteremic microorganisms (isolate numberN10) in the elderly (case patients) and those in the younger adults (control patients)

Bacteremic Isolate number P

isolates (% of resistance) value

Case Control

patients patients

Male	11 (78.6)	98 (45.4)	4.42 (1.20-12.27)	.02
Inappropriate empiric antibiotic therapy	9 (64.3)	150 (69.4)	0.79 (0.26-2.45)	.77
Bacteremic features
Community-acquired	8 (57.1)	169 (78.2)	0.37 (0.12-1.12)	.10
Polymicrobial	2 (14.3)	21 (9.7)	1.55 (0.32-7.39)	.63
Pittsburgh bacteremia score >= 4 points	8 (57.1)	23 (10.6)	11.19 (3.57-35.10	b .001
Major comorbidities
Malignancy	7 (50.0)	44 (20.4)	3.91 (1.30-11.73)	.02
Hypertension	3 (21.4)	72 (33.3)	0.55 (0.15-2.02)	.56
Diabetes mellitus	4 (28.6)	79 (36.6)	0.69 (0.21-2.29)	.78
Chronic renal insufficiency	2 (14.3)	36 (16.7)	0.83 (0.18-3.88)	1.30
Liver cirrhosis	2 (14.3)	35 (16.2)	0.86 (0.19-4.02)	1.00
Congestive heart failure	1 (7.1)	9 (4.2)	1.77 (0.21-15.05)	.47
Old stroke	1 (7.1)	7 (3.2)	2.30 (0.26-20.09)	.40
Major sources of bacteremia
Pneumonia	3 (21.4)	15 (6.9)	3.65 (0.92-14.53)	.09
Primary bacteremia	3 (21.4)	38 (17.6)	1.28 (0.34-4.80)	.72
Vascular catheter-related infections	2 (14.3)	9 (4.2)	3.83 (0.74-14.75)	.14
Urinary tract infections	1 (7.1)	75 (34.7)	0.15 (0.02-1.14)	.04
Liver abscess	1 (7.1)	11 (5.1)	1.43 (0.17-11.97)	.54
Biliary tract infections	1 (7.1)	13 (6.0)	1.20 (0.15-9.91)	.60
Laboratory data in the ED* Leukocyte N 9,000/mm3	8/14 (57.1)	137/211 (64.9)	0.72 (0.24-2.15)	.57
Platelet b 100,000/mm3	5/14 (35.7)	35/208 (16.8)	2.75 (0.87-8.69)	.14
Blood urea nitrogen N 20 mg/dL	11/14 (78.6)	89/206 (43.2)	4.82 (1.31-17.79)	.01
Serum creatinine N 1.5 mg/dL	6/14 (42.9)	56/210 (26.7)	2.06 (0.69-6.21)	.22
C-reactive protein N 100 mg/L	5/11 (45.5)	73/196 (37.2)	1.40 (0.41-4.76)	.75
Glucose N 200 mg/dL	1/14 (7.1)	51/174 (29.3)	0.19 (0.02-1.46)	.12
serum albumin b 3.5 g/dL	5/7 (71.4)	35/64 (54.7)	2.07 (0.37-11.48)	.46

therapy (OR 3.7, 95% CI 1.0-13.3; P = .049), and serum

Table 5 Univariate analysis of risk factors in the emergency department (ED) for 28-day mortality in 230 younger adults (control patients) with bacteremia

Variables Patient number (%)

Non-survivor n = 14

Survivor n = 216

Odds ratio (95% CI)

P

values

creatinine N 1.5 mg/dL (OR 3.4, 95% CI 1.1-13.3; P = .04),

were independently associated with the 28-day mortality. The respective impact of inappropriate empiric therapy on the outcome of the case patients and control patients was evidenced in the Kaplan-Meier survival curves (Fig.).

Discussion

Several investigations have demonstrated the clinical impact of a delayed administration of the first dose of appropriate antibiotics to patients with severe infections, such as bacteremia, severe sepsis [23], pneumonia [24], or meningitis [25]. However, to our knowledge, there is no study focusing on the elderly (aged >= 65) with bacteremia to concern the impact of appropriate empiric therapy by the comparisons with the younger adults (aged b 65) in the English literature. By the multivariate regression analysis, we found that inappropriate empiric antibiotic therapy was an independent prognostic factor in the elderly with bacteremia, and in contrast the survival

impact of appropriateness of empirical therapy was not evident in younger adults with bacteremia. Therefore, we emphasize the consideration of the patient’s age as one of the critical issues in choosing empiric antimicrobial therapy and the strategy of empirical therapy for bloodstream infections should be different at least for the elderly and those aged b 65.

The increase in mortality and morbidity due to infections in the elderly had been well recognized and associated with significant alterations in the adaptive humoral and cell- mediated immunity that occurs with aging [6,7]. Moreover, in the previous investigation focusing on Gram-negative bacillary bacteremia, mortality appears to be significantly increased in subgroup of the elderly, as comparison with younger patients [26]. In accordance with these studies, we also found that elderly patients with bacteremia had a worse outcome than younger adults with bacteremia and it may be related to high proportion of Gram-negative aerobes (75.3%) in our population.

Several investigations have described a relationship between inappropriate empiric antibiotic therapy and elderly patients with infections, which was usually due to a greater

Male	23	(67.6)	124 (48.8)	2.19	(1.02-4.86)	.03
Inappropriate empiric antibiotic therapy	15	(44.1)	70 (27.6)	3.76	(1.83-7.77)	.04
Bacteremic features
Community-acquired	15	(44.1)	200 (78.7)	0.21	(0.10-0.45)	b .001
Polymicrobial	3	(8.8)	25 (9.8)	0.88	(0.25-3.10)	1.00
Pittsburgh bacteremia score >= 4 points	25	(73.5)	40 (15.7)	14.86	(6.45-34.19)	b .001
Admitted to ICUs through the ED	12	(35.3)	33 (13.0)	3.65	(1.65-4.07)	.001
Major comorbidities
Hypertension	19	(55.9)	145 (57.1)	0.95	(0.46-1.96)	.89
Malignancy	13	(38.2)	48 (18.9)	2.65	(1.24-5.67)	.01
Diabetes mellitus	14	(41.2)	93 (36.6)	1.21	(0.58-2.51)	.60
Chronic renal insufficiency	8	(23.5)	39 (15.4)	1.69	(0.71-4.02)	.22
Coronary artery diseases	9	(26.5)	27 (10.6)	3.02	(1.28-7.15)	.02
Liver cirrhosis	5	(14.7)	27 (10.6)	1.45	(0.51-4.05)	.55
Congestive heart failure	3	(8.8)	23 (9.1)	0.97	(0.27-3.42)	1.00
Old stroke	3	(8.8)	23 (9.1)	0.97	(0.27-3.42)	1.00
Sources of bacteremia
Pneumonia	13	(38.2)	29 (11.4)	4.80	(2.17-10.60)	b .001
Primary bacteremia	9	(26.5)	35 (13.8)	2.25	(0.97-5.22)	.05
Skin and soft-tissue infections	4	(11.8)	18 (7.1)	1.74	(0.55-5.51)	.30
Urinary tract infections	3	(8.8)	103 (40.6)	0.14	(0.04-0.47)	b .001
Vascular catheter-related infections	2	(5.9)	5 (2.0)	3.11	(0.58-16.71)	.18
Liver abscess	1	(2.9)	14 (5.5)	0.51	(0.06-4.08)	1.00
Biliary tract infections	1	(2.9)	30 (11.8)	0.22	(0.03-1.71)	.14
Intraabdominal infections	0	(0)	10 (3.9)	0.96	(0.93-1.01)	.61
Laboratory data in the ED*
Leukocyte N 9,000/mm3	17/33	(51.5)	168/253 (66.4)	0.53	(0.25-1.11)	.09
Platelet b 100,000/mm3	8/33	(24.2)	40/250 (16.0)	1.68	(0.70-3.99)	.23
Blood urea nitrogen N 20 mg/dL	28/34	(82.4)	154/249 (61.8)	2.87	(1.14-7.21)	.01
Serum creatinine N 1.5 mg/dL	19/34	(55.9)	73/249 (29.3)	3.05	(1.47-6.33)	.002
C-reactive protein N 100 mg/L	18/32	(56.3)	80/233 (34.3)	2.45	(1.16-5.20)	.01
Glucose N 200 mg/dL	10/31	(32.3)	74/227 (32.6)	0.98	(0.44-2.19)	.97
Serum albumin b 3.5 g/dL	10/12	(83.3)	39/66 (59.1)	3.46	(0.70-17.07)	.19
ICUs = intensive care units.

diversity of pathogens in the elderly patients [7,27] or Atypical clinical presentations [7,13]. However, in our study about 70% of the elderly and younger adults with bloodstream infections both received appropriate empirical antimicrobial agents. Such a finding probably is related to two aspects. First, there is similarity in the species distribution and antibiograms of bacterial isolates from the elderly and younger adults. Second, the elderly more often came from long-term healthcare facilities and received more cefuroxime or aminopenicillin/?-lactamase inhibitors that had a wider spectrum of coverage for bacteremic pathogens than cefazolin that was the most common drug used in younger adults.

Table 6 Univariate analysis of the risk factors in the emergency department (ED) for 28-day mortality in 288 elderly patients (case patients) with bacteremia

Variables Patient number (%)

Non-survivor, n = 34

Survivor, n = 254

Odds ratio (95% CI)

P

values

Critical illness as indicated by a high Pittsburgh bacteremia score (>= 4 points) has been associated with a poor prognosis in patients with bacteremia due to various microorganisms [20,28-30]. In our study there less patients

with critical illness among young adults (13.5%) than the elderly (22.6%), and not surprisingly there was a lower 28- day mortality rate in the former than that in the latter. Moreover, based on our published data, the disadvantage of inappropriate antibiotic treatment was evident in bacteremic adults visiting the ED, and more obvious for critically ill patients [31]. With regard to younger adults, similar proportions of critical illness were present in patients with appropriate and inappropriate antibiotic therapy (19/159, 11.9% vs. 12/71, 16.9%; P = .3), reftecting that inappropriate antibiotic coverage in the ED would not affect the mortality rate.

For ED clinician, the major therapeutic strategy for individual with suspected infection was empirical antibiotic therapy [23]. That was the reason that a ED-based investigation was chosen in the current study. Focusing on the elderly visiting the ED, the empirical antibiotic therapy

Fig The Kaplan-Meier survival curves of younger adults (age b 65 years) (A, P = .67) and the elderly (age >= 65 years) (B, P = .049) empirically treated by inappropriate or appropriate antibiotics.

became difficult because a great diversity of pathogens depended on the setting of acquisition of infections (such as in the community, in a long-term care facility, or an acute- care hospital) and a practical difficulty in achieving microbiological procedure due to a anatomical alteration (such as lumbar puncture) [7]. Therefore, it was necessary to recognize the importance of empirical antibiotic therapy in an elderly subpopulation.

In addition to inappropriate antibiotic therapy, other prognostic factors were identified for the elderly with bacteremia by the multivariate analysis. Of them, two most significant host factors with an odds ratio N 10 were the Pittsburgh bacteremia score and underlying malignancy. The value of Pittsburgh bacteremia score in evaluating the severity of illness in bacteremia due to E. coli [28], K. pneumoniae [20,28], streptococci [29], or staphylococci [30], has been reported. These microorganisms were major bacteremic pathogens in the elderly of our population, and therefore it is not surprising to find that the Pittsburgh bacteremia score was an important predictor of mortality among them. In accordance with a published ED-based investigation of bacteremia [16], underling malignancy was independently related to mortality. In addition, a favorable prognostic factor was urinary tract infection as the source of bacteremia, and this finding was compatible to the results of two ED-based studies [16,31].

Limitations

We interpreted our data in light of several limitations inherent in the study design. First, to assess the outcome of the bacteremic patients treated and to avoid any potential ethical conftict, it was necessary to review the medical records and patients’ outcome retrospectively; therefore, there may have been errors due to confounding and bias.

Second, to adequately investigate the clinical effect of empiric antibiotic therapy in the ED, patents with bacteremia diagnosed prior to visiting the ED were excluded from the analysis, which may have led to some selection bias. However, only a small proportion of bacteremic patients were excluded, which probably only minimally inftuenced the results of current study. Third, we only focused on the empiric antibiotic therapy in the ED, where short lengths of stay and therapeutic duration in the ED were discovered and this “time” factor maybe mask other impacts of inappropriate empiric therapy. Finally, although we examined all bacter- emic patients who visited the ED over a long period of time, the study was conducted only in a single ED in a tertiary hospital. Therefore, our findings may not be generalizable to other populations.

Conclusions

For bacteremic adults visiting the ED, this study demonstrated that clinical impact of appropriateness of empirical antibiotic therapy on patients’ outcome in the elderly and younger adults varied. Further studies focusing on different strategies for Empiric treatment of bacteremia in the elderly and younger adults are warranted.

References

1. Bates DW, Pruess KE, Lee TH. How bad are bacteremia and sepsis? Outcomes in a cohort with suspected bacteremia. Arch Intern Med 1995;155:593-8.
2. Haug JB, Harthug S, Kalager T, Digranes A, Solberg CO. Bloodstream infections at a Norwegian university hospital, 1974-1979 and 1988- 1989: changing etiology, clinical features, and outcome. Clin Infect Dis 1994;19:246-56.
3. Leibovici L, Shraga I, Drucker M, Konigsberger H, Samra Z, Pitlik SD. The benefit of appropriate empirical antibiotic treatment in patients with bloodstream infection. J Intern Med 1998;244:379-86.
4. Harbarth S, Garbino J, Pugin J, Romand JA, Lew D, Pittet D. Inappropriate initial antimicrobial therapy and its effect on survival in a clinical trial of immunomodulating therapy for severe sepsis. Am J Med 2003;115:529-35.
5. Salomao R, Castelo Filho A, Pignatari AC, Wey SB. Nosocomial and community acquired bacteremia: variables associated with outcomes. Rev Paul Med 1993;111:456-61.
6. Grubeck-Loebenstein B, Wick G. The aging of the immune system. Adv Immunol 2002;80:243-84.
7. Gavazzi G, Krause KH. Ageing and infection. Lancet Infect Dis 2002;

2:659-66.

1. Opal SM, Girard TD, Ely EW. The immunopathogenesis of sepsis in elderly patients. Clin Infect Dis 2005;41(Suppl. 7):S504-12.
2. Martin GS, Mannino DM, Moss M. The effect of age on the development and outcome of adult sepsis. Crit Care Med 2006;34: 15-21.
3. Lesens O, Methlin C, Hansmann Y, Remy V, Martinot M, Bergin C, et al. Role of comorbidity in mortality related to Staphylococcus aureus bacteremia: a prospective study using the Charlson weighted index of comorbidity. Infect Control Hosp Epidemiol 2003;24: 890-6.
4. Strange GR, Chen EH, Sanders AB. Use of emergency departments by elderly patients: projections from a multicenter data base. Ann Emerg Med 1992;21:819-24.
5. Richardson JP. Bacteremia in the elderly. J Gen Intern Med 1993;8: 89-92.
6. Jarrett PG, Rockwood K, Carver D, Stolee P, Cosway S. Illness presentation in elderly patients. Arch Intern Med 1995;155:1060-4.
7. Yoshikawa TT. Perspective: aging and infectious diseases: past, present, and future. J Infect Dis 1997;176:1053-7.
8. Weiss SJ, Derlet R, Arndahl J, Ernst AA, Richards J, Fernandez- Frackelton M, et al. Estimating the degree of emergency department overcrowding in academic medical centers: results of the National ED Overcrowding Study (NEDOCS). Acad Emerg Med 2004;11:38-50.
9. Lin JN, Tsai YS, Lai CH, Chen YH, Tsai SS, Lin HL, et al. risk factors for mortality of bacteremic patients in the emergency department. Acad Emerg Med 2009;16:749-55.
10. Clinical and Laboratory Standards Institute. Performance standards for antimicrobial disk susceptibility tests, 17th informational supplement. document M100-S16. Wayne (Pa): CLSI; 2007.
11. Lee CC, Lin WJ, Shih HI, Wu CJ, Chen PL, Lee HC, et al. Clinical significance of potential contaminants in blood cultures among patients in a medical center. J Microbiol Immunol Infect 2007;40: 438-44.
12. Siegman-Igra Y, Fourer B, Orni-Wasserlauf R, Golan Y, Noy A, Schwartz D, et al. Reappraisal of community-acquired bacteremia: a proposal of a new classification for the spectrum of acquisition of bacteremia. Clin Infect Dis 2002;34:1431-9.
13. Paterson DL, Ko WC, Von Gottberg A, Mohapatra S, Casellas JM, Goossens H, et al. International prospective study of Klebsiella pneumoniae bacteremia: implications of extended-spectrum beta- lactamase production in Nosocomial infections. Ann Intern Med 2004;140:26-32.
14. Schellevis FG, van der Velden J, van de Lisdonk E, van Eijk JT, van Weel C. Comorbidity of chronic diseases in General practice. J Clin Epidemiol 1993;46:469-73.
15. Garner JS, Jarvis WR, Emori TG, Horan TC, Hughes JM. CDC definitions for nosocomial infections, 1988. Am J Infect Control 1988; 16:128-40.
16. Nguyen HB, Rivers EP, Abrahamian FM, Moran GJ, Abraham E, Trzeciak S, et al. Severe sepsis and septic shock: review of the literature and emergency department management guidelines. Ann Emerg Med 2006;48:28-54.
17. Pines JM, Morton MJ, Datner EM, Hollander JE. Systematic delays in antibiotic administration in the emergency department for adult patients admitted with pneumonia. Acad Emerg Med 2006;13:939-45.
18. Proulx N, Frechette D, Toye B, Chan J, Kravcik S. Delays in the administration of antibiotics are associated with mortality from adult acute Bacterial meningitis. QJM 2005;98:291-8.
19. McCue JD. Gram-negative bacillary bacteremia in the elderly: incidence, ecology, etiology, and mortality. J Am Geriatr Soc 1987;35:213-8.
20. Leibovici L, Pitlik SD, Konisberger H, Drucker M. Bloodstream infections in patients older than eighty years. Age Ageing 1993;22: 431-42.
21. Lee CI, Lee NY, Yan JJ, Lee HC, Ko NY, Chang CM, et al. Extended- spectrum beta-lactamase-producing phenotype signifies a poor prognosis for patients with cefpodoxime-resistant Escherichia coli or Klebsiella pneumoniae bacteremia. J Microbiol Immunol Infect 2009;42:303-9.
22. Yu VL, Chiou CC, Feldman C, Ortqvist A, Rello J, Morris AJ, et al. An international prospective study of pneumococcal bacteremia: correlation with in vitro resistance, antibiotics administered, and clinical outcome. Clin Infect Dis 2003;37:230-7.
23. Chang FY, Peacock Jr JE, Musher DM, Triplett P, MacDonald BB, Mylotte JM, et al. Staphylococcus aureus bacteremia: recurrence and the impact of antibiotic treatment in a prospective multicenter study. Medicine (Baltimore) 2003;82:333-9.
24. Lee CC, Lee CH, Chuang MC, Hong MY, Hsu HC, Ko WC. Impact of inappropriate empirical antibiotic therapy on outcome of bacteremic adults visiting the ED. Am J Emerg Med 2011 (in press).